Liver X receptor agonist inhibits proliferation of ovarian carcinoma cells stimulated by oxidized low density lipoprotein

Liver X receptor agonist inhibits proliferation of ovarian carcinoma cells stimulatedby oxidized low density lipoprotein

Daniel R. Scoles a,b,⁎, Xuan Xu a, Haimei Wang c, Hang Tran a, Barbie Taylor-Harding b,Andrew Li a,b, Beth Y. Karlan a,b

a Women's Cancer Research Institute and Division of Gynecologic Oncology, CSMC Burns and Allen Research Institute, Cedars-Sinai Medical Center, 8700 Beverly Boulevard,Los Angeles, CA, USAb Department of Obstetrics and Gynecology, David Geffen School of Medicine, University of California at Los Angeles, Los Angeles, CA, USAc Division of Cardiothoracic Surgery, CSMC Burns and Allen Research Institute, Cedars-Sinai Medical Center, 8700 Beverly Boulevard, Los Angeles, CA, USA

a b s t r a c ta r t i c l e i n f o

Article history:Received 17 April 2009Available online 24 October 2009

Keywords:LXR agonistTO901317Ovarian cancerOxidized LDLChemoresistanceStatinFluvastatin

Objectives. We previously observed an association between ovarian cancer outcome and statin use andhypothesized lipoproteins have direct effects on ovarian cancer proliferation. Here we investigate the directeffects of low density lipoprotein (LDL) and oxidized LDL (oxLDL) on proliferation and the inhibitory effectsof fluvastatin and a liver X receptor (LXR) agonist.

Methods. The effects of LDL, oxLDL, the LXR agonist TO901317, fluvastatin and cisplatin on cellularproliferation were determined using MTT assays. LXR pathway proteins were assayed by immunoblotting.Cytokine expression was determined by antibody array.

Results. Concentrations of oxLDL as small as 0.1 μg/ml stimulated CAOV3 and SKOV3 proliferation, whileLDL had no effect. TO901317 inhibited the proliferation of CAOV3, OVCAR3 and SKOV3 cells stimulated byoxLDL. Fluvastatin inhibited oxLDL mediated proliferation of CAOV3 and SKOV3. Cardiotrophin 1 (CT-1) wasmitogenic to CAOV3 and SKOV3, was induced by oxLDL, and was reversed by TO901317. OxLDL increasedcisplatin IC50s by 3.8 μM and N 60 μM for CAOV3 and SKOV3 cells, respectively. The LXR pathway proteinsCD36, LXR, and ABCA1 were expressed in eight ovarian carcinoma cell lines (A2780, CAOV3, CP70, CSOC882,ES2, OVCAR3, SKOV3).

Conclusions. OxLDL reduced ovarian carcinoma cell chemosensitivity and stimulated proliferation. Theseeffects were reversed by LXR agonist or fluvastatin. The LXR agonist also inhibited expression of the ovariancancer mitogen CT-1. These observations suggest a biologic mechanism for our clinical finding that ovariancancer survival is associated with statin use. Targeting LXR and statin use may have a therapeutic role inovarian cancer.

© 2009 Elsevier Inc. All rights reserved.

Introduction

The association between serum cholesterol and cancer risk andsurvival is unclear. Attempts to correlate risk or outcome of numerouscancers with serum low density lipoprotein (LDL) cholesterol haveyielded somewhat controversial results with evidence pointingtowards improved outcome in patients with high serum cholesterol[1–3]. One study also demonstrated that women on a diet high incholesterol had no increased risk of epithelial ovarian cancer [4].However, recently it was reported that the levels of oxidized LDL

(oxLDL) in the serum of ovarian cancer patients were positivelyassociated with patient outcome [5]. Our group has also recentlyreported that epithelial ovarian cancer patients on statins hadimproved survival [6].

Levels of oxLDL are in part regulated by the available oxidizableLDL precursor and may be altered by cholesterol lowering drugs suchas statins. Cholesterol homeostatis is maintained in part by cellsexpressing scavenger receptors (SRA and CD36) that internalizeoxLDL which are then converted to oxysterol ligands of the nuclearliver X receptors α and β (LXRα and LXRβ), heterodimers of theretinoid X receptor (RXR) [7]. Activated LXR/RXR heterodimersactivate target genes possessing the LXR element (LXRE) includingthe ATP binding cassette transporters ABCA1 and ABCG1 leading tocholesterol efflux to HDL or cholesterol excretion via bilial andintestinal cells [7]. LXR activation is also associated with increases inthe expression of numerous proinflammatory cytokines which alsopromote cellular proliferation [7]. While statin therapy can lower

Gynecologic Oncology 116 (2010) 109–116

Abbreviations: LXR, liver X receptor; LDL, low density lipoprotein; oxLDL, oxidizedLDL.⁎ Corresponding author. Department of Neurology, University of Utah, 175 North

Medical Center Drive East, 5th Floor, Salt Lake City, UT, 84132, USA. Fax: +1 801 5816707.

E-mail address: [email protected] (D.R. Scoles).

0090-8258/$ – see front matter © 2009 Elsevier Inc. All rights reserved.doi:10.1016/j.ygyno.2009.09.034

Contents lists available at ScienceDirect

Gynecologic Oncology

j ourna l homepage: www.e lsev ie r.com/ locate /ygyno

serum oxLDL [8-11], statins also have lipid independent functions onoxidative pathways including regulation of scavenger receptors. Inmonocytes and foam cells, statins can alter scavenger receptorexpression via regulation of PPAR, NF-κB, Rho and Ras [12], butthese statin functions have not been studied in cancer cells.

One potential means for treating oxLDL related diseases ismodulation of LXR by synthetic LXR agonists. Presently three suchsynthetic LXR ligands have been developed including TO901317,GW3965 and N,N-dimethyl-3β-hydroxycholenamide (DMHCA). LXRagonists upregulate expression of the ATP binding cassette transpor-ters and cholesterol efflux while simultaneously downregulatingoxysterol mediated activation of proinflammatory cytokines [13,14].In mouse models of atherosclerosis LXR agonists reduce bothinflammation in atherosclerotic plaques and serum cholesterol [15].LXR agonists also inhibit the proliferation of endothelial cells [16].oxLDL and LXR agonists have been little studied on cells other thanendothelial, monocytes and macrophage foam cells in the context ofcardiovascular diseases, but one study demonstrated oxLDL is amitogen to cultured human fibroblasts [17].

In the present study we determined that ovarian carcinoma cellspossess CD36 scavenger receptor and are stimulated to proliferate byoxLDL. The LXR agonist TO901317 and fluvastatin reversed oxLDLmediated proliferation.We also demonstrated that oxLDL reduced thesensitivity of ovarian carcinoma cells to cisplatin. Our studydemonstrated that increases in oxidized LDL cholesterol maynegatively impact ovarian cancer outcome and suggests that LXRligands and statins may be an effective strategy for treating ovariancancer patients.

Materials and methods

Tissue culture

CAOV3, ES2, OVCAR3, PA1, and SKOV3 were cultured as recom-mended by American Type Culture Collection. CSOC882 and CSOC909were cultured as previously described [18]. A2780 and CP70 werecultured in RPMI 1640 and 10% fetal calf serum (FCS) supplementedwith 2 mM L-glutamine and 0.2 U/ml insulin. OVCA432 was culturedin modified Eagle's medium (MEM) and 10% FBS+2mM L-glutamine.All reagents were purchased from GIBCO.

RNA interference

RNA interference (RNAi) was accomplished by suspensiontransfection of CAOV3 cells using Metafectene (Biontex Laborato-ries) and 30 nM of the indicated siRNA oligos. When two siRNAswere used they were 15 nM each. Transfected cells were plated in96 well plates in quadruplicate, 10,000 cells/well. siRNAs includedLXRα (sc-38828), LXRβ (sc-45316) and Control siRNA-A (sc-37007)(Santa Cruz Biotechnology). The next day cells were serum starved4 h and treated with the indicated amounts of TO901317 overnight.Cell abundances were determined by MTT assay as describedbelow.

Immunoblotting

Proteins were separated on precast polyacrylamide gels (Bio-Rad),transferred to Hybond ECL (Amersham), and detected by enzymaticchemiluminescence (ECL) (Amersham). Antibodies included rabbitanti-CD36 scavenger receptor (Santa Cruz #sc-9154, used at 2 μg/ml),goat anti-LXRα/β (Novus Biologicals #NB100-1465, used at 2 μg/ml),rabbit anti-ABCA1 (Novus Biologicals #NB400-105, used at 2 μg/ml),rabbit anti-NAP2 (Abcam Inc. #ab9554, used at 0.25 μg/ml), mouseanti-actin monoclonal antibody (Sigma-Aldrich #AC-40, used at1:1000), and rabbit anti-CT1 (Abcam #ab9837, used at 0.5 μg/ml).Secondary antibodies conjugated to horseradish peroxidase were

purchased from Jackson ImmunoResearch Laboratories, Inc. Imageswere processed densitometrically using ImageJ [19].

Proliferation assays

Ovarian carcinoma cell lines were seeded in 96 well plates at adensity of 3000 cells per well. The next day cells were treated inreduced serum media (0.1% FBS) for 4 h then were treated with theindicated concentrations of cisplatin (Sigma), oxLDL (BiomedicalTechnologies), TO901317 (Cayman Chemicals), or CT-1 (CellSciences) for 24 or 48 h. Cell abundances were determined usingthe CellTiter kit (Promega). MTT concentration was determined byOD490.

Antibody array analysis

SKOV3 cells were cultured in 10 cm dishes overnight. The next daycells were serum starved for 4 h then were treated with combinationsof oxLDL (25 μg/ml), 25 ng/ml TO901317 and diluent as indicatedfor 12 h. Media were then collected and filtered to exclude cells.This conditioned media was then used without dilution to probe ahuman cytokine antibody array (RayBiotech, Inc. #AAH-CYT-8)following the vendor's recommended protocol using kit reagents. Theprocess involved detection of cytokines from media bound toantibody arrays using a cocktail of corresponding biotin conjugatedanti-cytokine antibodies subsequently labeled by HRP-conjugatedstreptavidin, revealed by enzymatic chemiluminescence. Films ex-posed to ECL were photographed at high resolution using a UVPbioimaging system (UVP, LLC). Images were then processed usingImageJ [19]. Fold inductions were calculated relative to arrayedcontrol proteins.

Results

oxLDL stimulated proliferation of ovarian carcinoma cell lines, but LDLhad no effect

We determined the effects of LDL and ox-LDL on proliferation ofSKOV3 and CAOV3 ovarian carcinoma cell lines by MTT assays.Increasing doses of LDL did not significantly alter the growth of eitherof these cell lines (Fig. 1A). However, the proliferation of each of thesecell lines was significantly increased upon treatment with oxLDL (Fig.1B). Increased proliferation was also evident visually by examinationof cells treated for 24 h with oxLDL by standard phase microscopy(Fig. 1C).

LXR pathway proteins are expressed in ovarian carcinoma cell lines

We compared the expression of the LXR pathway proteins in sevenovarian carcinoma cell lines including CD36 scavenger receptor, LXRα/β, and ABCA1. The cell lines included A2780, CAOV3, CP70,CSOC882, ES2, OVCAR3, SKOV3. Each of these proteins were presentin ovarian carcinoma cell lines but in variable amounts (Figs. 1D andE).We concluded from these results that the CD36!LXR pathway isintact in most ovarian carcinoma cell lines.

LXR agonist TO901317 reversed oxLDL mediated proliferation

We hypothesized that activating the LXR pathway would inhibitproliferation of ovarian carcinoma cell lines expressing LXR pathwayproteins. Treatment of each of OVCAR3, SKOV3, and CAOV3 withincreasing doses of TO901317 inhibited proliferation in a dose-dependent manner (Pb0.01, t-test). These data demonstrate that LXRagonist can inhibit proliferation without oxLDL stimulation (Fig. 2A).TO901317 treatment also completely reversed the proliferation ofeach of these three cell lines stimulated by simultaneous oxLDL

110 D.R. Scoles et al. / Gynecologic Oncology 116 (2010) 109–116

treatment (Fig. 2B). The proliferation of CAOV3 cells treated withsiRNAs against LXRα, LXRβ or a combination of bothwas not inhibitedby TO901317. Slopes were not significantly different from zero(PN0.05) unless a control siRNA was used (P=0.05; Fig. 2C).Additionally, any LXR siRNA significantly inhibited proliferationcompared to control siRNA at any dose of TO901317 (Fig. 2C). Notethat Western blotting was used to show siRNA treatment achievedpartial reduction of LXR protein.

OxLDL reduced sensitivity to fluvistatin

Treatments of oxLDL significantly reduced the sensitivity of each ofCAOV3 and SKOV3 to fluvastatin. For CAOV3, the fluvistatin IC50swere 248 μM (!oxLDL) and 341 μM (+oxLDL), and oxLDL treatmentresulted in an increase of the fluvistatin IC50 by 93 μM (Fig. 3A). Thecurves corresponding to CAOV3 treatments with and without oxLDLwere significantly different by two-way ANOVA (Pb0.001). ForSKOV3, the fluvistatin IC50s were 134 μM (!oxLDL) and 373 μM(+oxLDL), and oxLDL treatment resulted in an increase of thefluvistatin IC50 by 239 μM (Fig. 3B). The curves corresponding to

SKOV3 treatments with and without oxLDL were significantlydifferent by two-way ANOVA (Pb0.0001).

Cytokine array screening

Screening of a cytokine antibody array of 54 cytokines with themedia of cultured SKOV3 cells conditioned with diluent, oxLDL, oroxLDL and TO901317 revealed numerous changes in the expression ofvarious cytokines. Supplementary Table 1S presents average foldchanges of the individual cytokines on the array determined byImageJ analysis. With 2-fold as a threshold for induction, somecytokines were induced by oxLDL but not reversed by TO901317 (IL-1ra, IL-2, NAP-2), while others were not induced by oxLDL butreductions were observed upon treatment with TO901317 (IL-2Rβand IL-2Rγ). For cardiotropin-1 (CT-1) we observed both an average2.3 fold induction by oxLDL and an average 2.5 fold reversal byTO901317 by ImageJ analysis, that was also clearly evident by visualinspection (Fig. 4A). We used immunoblotting to verify that CT-1expression was indeed induced upon treatment with oxLDL in bothSKOV3 cells and also CAOV3 cells (Figs. 4B, C).

Fig. 1. Effect of LDLs on proliferation and presence of LXR pathway proteins. (A) Increasing doses of LDL did not alter proliferation of SKOV3 or CAOV3 cells. (B) Increasing doses ofoxLDL increased SKOV3 and CAOV3 cell proliferation. In both A and B, cells were treated for 24 h and abundances were determined by MTT assays. Values shown are means±SDfrom 4 replicates. Doses used were 0, 5, 15, 25, 50 and 100 μg/ml LDL (A) and 0, 0.05, 0.1, 0.5, 1, 5, 25, 50, 100, and 200 μg/ml oxLDL, coded by the addition of 0.02 before the logtransformation (B). (C) Phase contrast image of SKOV3 cells treated with and without 25 μg/ml oxLDL for 24 h. (D) Immunoblots showing the expression of LXR pathway proteinsCD36, LXRα/β, and ABCA1. (E) Densitometric quantification of the abundances of LXR pathway proteins in the immunoblots shown in D, relative to actin.

111D.R. Scoles et al. / Gynecologic Oncology 116 (2010) 109–116

Cardiotropin-1 stimulated ovarian carcinoma cell proliferation

We determined whether CT-1 could induce proliferation ofovarian carcinoma cells since CT-1 was the only cytokine in ouranalysis whose expressionwas stimulated by oxLDL greater than two-fold that was reversed by TO901317. We treated both SKOV3 andCAOV3 cells with increasing doses of CT-1 demonstrating enhancedproliferation for both cell lines (Fig. 4D). The increases in proliferationwere significantly different by one-way ANOVA (Pb0.0001 andPb0.05 for CAOV3 and SKOV3, respectively). The highest dose of CT-1 used, 160 ng/ml, increased proliferation in a 24-h period by 1.34

fold for CAOV3 and 1.15 fold for SKOV3 compared to diluent treatedcontrols.

OxLDL reduced cisplatin chemosensitivity

Treatments with oxLDL significantly reduced the chemosensitivityof each of CAOV3 or SKOV3. For CAOV3, the cisplatin IC50s were4.5 μM (!oxLDL) and 7.4 μM (+oxLDL), and oxLDL treatmentresulted in an increase of the cisplatin IC50 by 2.9 μM (Fig. 5A). ForSKOV3, the cisplatin IC50s were 42 μM (!oxLDL) and 111 μM(!oxLDL, determined by extrapolation since cells treated with the

Fig. 2. LXR agonist reversed proliferation mediated by oxLDL. (A) Increasing doses of TO901317 inhibited the proliferation of OVCAR3, SKOV3, and CAOV3. (B) Paired treatments ofdiluent, 25 μg/ml oxLDL, or 25 μg/ml oxLDL plus 25 ng/ml TO901317 demonstrated that TO901317 fully reversed the proliferation of OVCAR3, SKOV3, and CAOV3 mediated byoxLDL. Cells were treated for 24 h and abundances were determined byMTT assays. (C) Increasing doses of TO901317 did not inhibit proliferation of CAOV3 cells treated with siRNAsagainst LXRα or LXRβ. Values shown are means±SD from 3 replicates (A and B) or 4 replicates (C).

Fig. 3. Fluvastatin inhibited the proliferation of ovarian carcinoma cells and oxLDL reduced the fluvastatin effect. We treated CAOV3 cells (A) and SKOV3 cells (B) with or without25 μg/ml oxLDL and increasing doses of fluvasatin for 24 h and determined cell abundances by MTT assays. OxLDL increased the IC50 by 92 μM fluvastatin for CAOV3 (from 248 to341) and 239 μM fluvastatin for SKOV3 (from 134 to 373). Doses used were 0, 7.8, 15.6, 31.2, 62.5, 125, 250, 500 and 1000 μM fluvastatin, coded by the addition of 1 before the logtransformation. Values shown are means±SD from three replicates.

112 D.R. Scoles et al. / Gynecologic Oncology 116 (2010) 109–116

highest cisplatin dose (100 μM)hadnot quite reached a50% reduction)and oxLDL treatment resulted in an increase of the cisplatin IC50 by69 μM (Fig. 5B). For each of SKOV3 and CAOV3, the curvescorresponding to treatments with and without oxLDL were signifi-cantly different by two-way ANOVA (Pb0.0001 for each cell line).

Discussion

Ovarian cancer outcome may be influenced by genetic, epige-netic, and clinical determinates as well as metabolic conditions that

alter cancer proliferation. Recently we demonstrated that ovariancancer patients taking statins had improved survival [6]. In addition,a recent meta-analysis of 37,248 individuals found reduced cancerincidence for statin users [20]. While not all studies havedemonstrated positive associations between serum cholesterollevels and outcome in solid malignancies [1–4], one studydemonstrated a direct correlation between serum oxLDL andovarian cancer outcome [5]. Prompted by these clinical observations,the present study sought to investigate the hypothesis thatoxLDL, TO901317 and fluvastatin may have direct effects on the

Fig. 4. Cardiotropin 1 secretion was increased by oxLDL and reversed by LXR agonist. (A) We screened for changes in the secretion of 54 proinflammatory cytokines by probing anantibody array with the media of SKOV3 cells treated with 25 μg/ml oxLDL, 25 μg/ml oxLDL and 25 ng/ml TO901317, or diluent for 24 h. Densitometric analysis demonstrated thatCT-1 was induced by oxLDL and reversed by TO901317 (indicated by the box). SKOV3 (B) cells or CAOV3 cells (C) were treated with or without oxLDL and increasing doses ofTO901317 and determined CT-1 abundance relative to actin by immunoblotting. For both cell types CT-1 expression was induced by oxLDL treatment, and the highest doses ofTO901317 suppressed the CT-1 expressionmediated by oxLDL. (D) CT-1 enhanced the proliferation of SKOV3 and CAOV3. Cells were treated with diluent alone or increasing doses ofCT-1 for 24 h and cell abundances were determined by MTT assays. Doses used were 0, 5, 10, 20, 40, 80, 160 ng/ml CT-1, coded by the addition of 2 before the log transformation.Values shown are means±SD from three replicates. Identities of all cytokines on the array can be found in Supplemental Table 2.

113D.R. Scoles et al. / Gynecologic Oncology 116 (2010) 109–116

proliferation of ovarian cancer cells, and expression of proinflam-matory cytokines.

Oxidized LDLs are mitogenic to ovarian cancer cells

Oxidized LDL stimulated ovarian carcinoma cells to proliferatewhile generic LDL did not. We investigated SKOV3 and CAOV3 ovariancancer cell lines for their proliferative response to LDL and observedno significant changes in cellular growth over the full range of dosestested. Contrary to expectations, all doses of LDL elicited decreases inproliferation for both cell lines compared to no treatment at all, but nofurther decreases were observed with higher doses (Fig. 1A).However, oxidized LDL potently stimulated both the SKOV3 andCAOV3 cell lines to proliferate (Fig. 1B). Our data support thehypothesis that elevated oxLDL may worsen ovarian cancer outcomeby stimulating cancer proliferation and are consistent with resultsfrom a clinical study showing significantly greater serum oxLDL levelsin ovarian cancer patients compared to a healthy controls [5]. Ourinitial findings on LDL and oxLDL mediated proliferation directed ourattention to modulation of the liver X receptor pathway as a regulatorof oxLDL.

LXR agonists reverse oxLDL mitogenesis of ovarian cancer cells

LXR agonists activate LXR more potently than oxysterolsresulting in elevated cholesterol efflux by genes possessing LXRelements (LXREs), and reduction of the expression of proinflamma-tory cytokines by genes lacking LXREs (Reviewed by [14]). Becausesome studies have shown lipid independent actions of LXR agonists[21,22] we investigated whether TO901317 could inhibit prolifera-tion of ovarian cancer cells both in the absence and presence ofoxLDL stimulation. We demonstrated that TO901317 alone signifi-cantly reduced cell abundances in proliferation assays, and thatTO901317 fully reversed the growth stimulated by oxLDL. Immuno-blotting demonstrated that each of seven ovarian cancer cell linesexpressed LXR pathway proteins CD36, LXR, and ABCA1, and that thethree that we further investigated, OVCAR3, SKOV3, and CAOV3,

expressed the highest levels of LXR. We also demonstrated that LXRagonist did not inhibit proliferation of CAOV3 cells treated with eachof three combinations of LXR siRNAs, and the effects of LXR siRNAswere significantly different from that observed using a controlsiRNA. Indeed when both LXRα and LXRβ siRNAs were used incombination a slight increase of proliferation was observed whichmight be due to multiple actions of the agonist. While weconsistently found that the potency of TO901317 was reducedwhen transfecting siRNAs, we also observed that treatment ofCAOV3 cells with any combination of LXR siRNAs significantlyreduced proliferation compared to the control. This observation isentirely consistent with LXR as a mediator of proliferation byactivation of proinflammatory cytokines, which is suppressed by LXRactivation by agonists. These data demonstrate that the observedTO901317 effects were mediated by LXR. We are aware of only oneother group of investigators that demonstrated the effects of anyLXR agonist on the proliferation of cancer cells. Their studiesdemonstrated that treatment of LNCaP prostate cancer cells withTO901317 significantly reduced proliferation of cells cultured invitro as well as in xenografts [23,24].

Fluvastatin inhibits proliferation of ovarian cancer cells

Statins are pleiotrophic compounds with inhibitory actions oncellular proliferation that are unrelated to cholesterol lowering. Wetested only fluvastatin on platinum-sensitive and -resistant ovariancancer cell lines to determine whether statins might have directinhibitory effects on the proliferation of ovarian cancer cells. Ourstudy demonstrated that fluvastatin inhibited the proliferation of bothCAOV3 and SKOV3 ovarian carcinoma cell lines, and was uniqueamong other studies in that we demonstrated fluvastatin could alsoinhibit proliferation mediated by oxLDL. Simvastatin, lovastatin andmevastatin were all effective for reducing proliferation of TOV112Dovarian carcinoma cells in the only other study to investigate anovarian carcinoma cell line [25]. There are now a number of studiesthat have investigated statin effects on the proliferation of othercancer cell types demonstrating statins are generally effective for

Fig. 5. OxLDL decreased ovarian carcinoma cell chemosensitivity. We treated CAOV3 cells (A) and SKOV3 cells (B) with or without 25 μg/ml oxLDL and increasing doses of cisplatinfor 24 h and determined cell abundances by MTT assays. OxLDL increased the IC50 by 2.9 μM cisplatin for CAOV3 (from 4.5 to 7.4) and 69 μM cisplatin for SKOV3 (from 42 to 111). Allvalues were coded by the addition of 1.0 before the log transformation. Doses used were 0, 2.5, 5, 10, 20 and 40 μM cisplatin (A) or 0, 2.5, 5, 10, 20, 40 and 100 μM cisplatin (B). Valuesshown are means±SD from 3 replicates (A) or 6 replicates (B).

114 D.R. Scoles et al. / Gynecologic Oncology 116 (2010) 109–116

reducing cancer proliferation [26]. For breast cancer cells in culture,Cambell et al. [27] showed that lipophilic statins were most potentand that these statins best inhibited cells with constitutively activeMAP-kinase pathways.

Cardiotrophin 1 expression was induced by oxLDL and reversed by LXRagonist

Our study also identified cardiotrophin 1 as a cytokine that wasinduced by oxLDL but reversed by TO901317. We made thisobservation by screening a cytokine antibody array with the mediaof SKOV3 cells treated with or without oxLDL and TO901317. Thepurpose of this experiment was to demonstrate in principle that theLXR agonist could reverse the expression of proinflammatorycytokines for proliferation induced by oxLDL treatment. Of all thecytokines on the array, CT-1 was the only one induced by oxLDL bygreater than 2 fold and fully reversed by TO901317 (see Supplemen-tary Tables 1S and 2S for the complete list).

CT-1 is a member of the IL-6 superfamily and activates the IL-6receptor gp130 [28]. CT-1 expression is anti-apoptotic and loss ofCT-1 in cardiomyocytes is associated with increased cell deathrelated to heart failure (Lopez et al., 2007). CT-1 has also beencharacterized as a hepatocyte stimulating factor and has been usedtherapeutically in cirrhotic rats to stimulate angiogenesis and liverregeneration in a manner not requiring gp130 expression suggest-ing that another CT-1 receptor may exist [29]. CT-1 is also ex-pressed in adipose tissue and upregulated in metabolic syndrome[33]. Therefore, elevated CT-1 levels due to obesity could contri-bute to reduced ovarian cancer survival, as previously observed byour group. Since PPAR, NFkB and Ras support CD36 expression [12],CT-1 expression might also promote oxLDL signaling as PPAR andNFkB were induced in various cell types including adipocytes, mo-nocytes and hepatocytes [30,31] and gp130 activates Ras in numer-ous cell types [32].

oxLDL reduces chemosensitivity of ovarian cancer cells

The observation that oxLDL potently stimulated the proliferationof ovarian cancer cells raised the question whether oxLDL could alterthe sensitivity of ovarian cancer cells to chemotherapy. We tested thishypothesis by treating both CAOV3 and SKOV3 cells with or withoutoxLDL and increasing doses of cisplatin and observed that oxLDLtreatment of both cell lines significantly reduced the cisplatin IC50s.Our study suggests that therapies that reduce oxLDL effects mighthave a role in augmenting chemoresponse and/or overcomingchemoresistance.

LXR agonists and statins as therapeutics for ovarian cancer

Our data suggest further investigation into the use of statins andLXR agonists as adjunctive therapies for ovarian cancer is merited.Statins may be particularly effective since they might reduce thenegative effects of high serum cholesterol, while simultaneouslyinhibiting proliferation of cancer cells stimulated by oxLDL and othermitogens. Like statins, LXR agonists are multifunctional, enhancingcholesterol efflux stimulated by oxLDL while simultaneously reducingthe expression of proinflammatory cytokines. We have now demon-strated that both fluvastatin and TO901317 are direct inhibitors ofovarian cancer cell proliferation in the absence of oxLDL and that eachcan reverse ovarian cancer cell proliferation mediated by oxLDL.Furthermore, TO901317 reduces ovarian cancer expression of themitogen CT-1 that is stimulated by oxLDL. In Fig. 6, we present amodel for how LXR agonists and statins might have a role as adjuvanttherapies or as primary or secondary prevention strategies for ovariancancer. Further investigation of statins, oxLDL, and the LXR pathway inovarian cancer are warranted.

Conflict of interest statementThe authors have no conflicts of interest to declare.

Acknowledgments

This work was supported by contributions from the L&S MilkenFoundation and an American Cancer Society California Division EarlyDetection Professorship to BYK.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.ygyno.2009.09.034.

References

[1] Tanne JH. Meta-analysis says low LDL cholesterol may be associated with greaterrisk of cancer. BMJ 2007;335:177.

[2] Chang AK, Barrett-Connor E, Edelstein S. Low plasma cholesterol predicts anincreased risk of lung cancer in elderly women. Prev Med 1995;24:557–62.

[3] Fiorenza AM, Branchi A, Sommariva D. Serum lipoprotein profile in patients withcancer. A comparison with non-cancer subjects. Int J Clin Lab Res 2000;30:141–5.

[4] Genkinger JM, Hunter DJ, Spiegelman D, et al. A pooled analysis of 12 cohortstudies of dietary fat, cholesterol and egg intake and ovarian cancer. Cancer CausesControl 2006;17:273–85.

[5] Delimaris I, Faviou E, Antonakos G, Stathopoulou E, Zachari A, Dionyssiou-AsteriouA. Oxidized LDL, serum oxidizability and serum lipid levels in patients with breastor ovarian cancer. Clin Biochem 2007;40:1129–34.

[6] Elmore RG, Ioffe Y, Scoles DR, Karlan BY, Li AJ. Impact of statin therapy on survivalin epithelial ovarian cancer. Gynecol Oncol 2008;111:102–5.

[7] Jamroz-Wisniewska A, Wojcicka G, Horoszewicz K, Beltowski J. Liver X receptors(LXRs). Part II: non-lipid effects, role in pathology, and therapeutic implications.Postepy Hig Med Dosw (Online) 2007;61:760–85.

[8] Hogue JC, Lamarche B, Tremblay AJ, Bergeron J, Gagne C, Couture P. Differentialeffect of atorvastatin and fenofibrate on plasma oxidized low-density lipoprotein,inflammation markers, and cell adhesion molecules in patients with type 2diabetes mellitus. Metabolism 2008;57:380–6.

Fig. 6.Model for the actions of statins and LXR agonist on cancer proliferation. Oxysterolmetabolites of oxLDL accumulate upon oxLDL internalization by CD36 scavengerreceptor. Oxysterols bind LXR resulting in both the activation of LXRE containing genesstimulating cholesterol homeostasis and efflux, and the expression of cytokinespromoting inflammation, proliferation and angiogenesis. Revving up of LXR by LXRagonists simultaneously enhances cholesterol efflux by activation of LXRE-containinggenes, and inhibits the expression of proinflammatory cytokines lacking LXREs.Pleiotrophic statins may be therapeutic for ovarian cancer by both lowering hepaticLDL cholesterol while inhibiting cancer growth pathways which based on thecardiovascular literature may include Rho, Rac, PPAR and NFkB which can also promoteCD36 expression.

115D.R. Scoles et al. / Gynecologic Oncology 116 (2010) 109–116

[9] Inami S, Okamatsu K, TakanoM, et al. Effects of statins on circulating oxidized low-density lipoprotein in patients with hypercholesterolemia. Jpn Heart J 2004;45:969–75.

[10] Resch U, Tatzber F, Budinsky A, Sinzinger H. Reduction of oxidative stress andmodulation of autoantibodies against modified low-density lipoprotein afterrosuvastatin therapy. Br J Clin Pharmacol 2006;61:262–74.

[11] van Tits LJ, van Himbergen TM, Lemmers HL, de Graaf J, Stalenhoef AF. Proportionof oxidized LDL relative to plasma apolipoprotein B does not change during statintherapy in patients with heterozygous familial hypercholesterolemia. Atheroscle-rosis 2006;185:307–12.

[12] Hofnagel O, Luechtenborg B, Weissen-Plenz G, Robenek H. Statins and foam cellformation: impact on LDL oxidation and uptake of oxidized lipoproteins viascavenger receptors. Biochim Biophys Acta 2007;1771:1117–24.

[13] Joseph SB, Castrillo A, Laffitte BA, Mangelsdorf DJ, Tontonoz P. Reciprocalregulation of inflammation and lipid metabolism by liver X receptors. Nat Med2003;9:213–9.

[14] Freeman MW, Moore KJ. eLiXiRs for restraining inflammation. Nat Med 2003;9:168–9.

[15] Joseph SB, McKilligin E, Pei L, et al. Synthetic LXR ligand inhibits the developmentof atherosclerosis in mice. Proc Natl Acad Sci U S A 2002;99:7604–9.

[16] Blaschke F, Leppanen O, Takata Y, et al. Liver X receptor agonists suppress vascularsmooth muscle cell proliferation and inhibit neointima formation in balloon-injured rat carotid arteries. Circ Res 2004;95:e110–23.

[17] Zettler ME, Prociuk MA, Austria JA, Massaeli H, Zhong G, Pierce GN. OxLDLstimulates cell proliferation through a general induction of cell cycle proteins. AmJ Physiol Heart Circ Physiol 2003;284:H644–53.

[18] Scoles DR, Pavelka J, Cass I, et al. Characterization of CSOC 882, a novelimmortalized ovarian cancer cell line expressing EGFR, HER2, and activatedAKT. Gynecol Oncol 2007;104:120–8.

[19] Rasband WS. ImageJ, National Institutes of Health, Bethesda, Maryland, USA. In:http://rsb.info.nih.gov/ij/; 2004.

[20] Farwell WR, Scranton RE, Lawler EV, et al. The association between statins andcancer incidence in a veterans population. J Natl Cancer Inst 2008;100:134–9.

[21] Czech C, Burns MP, Vardanian L, et al. Cholesterol independent effect of LXRagonist TO-901317 on gamma-secretase. J Neurochem 2007;101:929–36.

[22] Chuu CP, Kokontis JM, Hiipakka RA, Liao S. Modulation of liver X receptor signalingas novel therapy for prostate cancer. J Biomed Sci 2007;14:543–53.

[23] Chuu CP, Hiipakka RA, Kokontis JM, Fukuchi J, Chen RY, Liao S. Inhibition of tumorgrowth and progression of LNCaP prostate cancer cells in athymic mice byandrogen and liver X receptor agonist. Cancer Res 2006;66:6482–6.

[24] Chuu CP, Chen RY, Hiipakka RA, et al. The liver X receptor agonist T0901317 acts asandrogen receptor antagonist in human prostate cancer cells. Biochem BiophysRes Commun 2007;357:341–6.

[25] Gauthaman K, Richards M, Wong J, Bongso A. Comparative evaluation of theeffects of statins on human stem and cancer cells in vitro. Reprod Biomed Online2007;15:566–81.

[26] Sassano A, Platanias LC. Statins in tumor suppression. Cancer Lett 2008;260:11–9.

[27] Campbell MJ, Esserman LJ, Zhou Y, et al. Breast cancer growth prevention bystatins. Cancer Res 2006;66:8707–14.

[28] Pennica D, Wood WI, Chien KR. Cardiotrophin-1: a multifunctional cytokine thatsignals via LIF receptor-gp 130 dependent pathways. Cytokine Growth Factor Rev1996;7:81–91.

[29] Yang ZF, Lau CK, Ngai P, et al. Cardiotrophin-1 enhances regeneration of cirrhoticliver remnant after hepatectomy through promotion of angiogenesis and cellproliferation. Liver Int 2008;28:622–31.

[30] Fritzenwanger M, Meusel K, Foerster M, Kuethe F, Krack A, Figulla HR.Cardiotrophin-1 induces interleukin-6 synthesis in human monocytes. Cytokine2007;38:137–44.

[31] Zvonic S, Hogan JC, Arbour-Reily P, Mynatt RL, Stephens JM. Effects ofcardiotrophin on adipocytes. J Biol Chem 2004;279:47572–9.

[32] Hirano T. Interleukin 6 and its receptor: ten years later. Int Rev Immunol 1998;16:249–84.

[33] Natal C, Fortuño MA, Restituto P, et al. Cardiotrophin-1 is expressed in adiposetissue and upregulated in the metabolic syndrome. Am J Physiol Endocrinol Metab2008;294:E52–60.

116 D.R. Scoles et al. / Gynecologic Oncology 116 (2010) 109–116